Who should own what AI produces?

The role of AI in media production raises the question of ownership.

An article by Kimberly Schnur and Paul Heß

Everyone is talking about artificial intelligence (AI). LinkedIn is full of self-proclaimed experts on the subject, YouTube is promoting get-rich-quick schemes using ChatGPT, and discussions about impending doom scenarios are dominating traditional media channels. Whilst this is happening, a real, tangible concern regarding AI is emerging in various fields: the question of ownership of media products produced by AI. The writers’ and actors’ unions in the United States, for example, were – and still are, at the time of writing this post – on strike over this issue. Questions about who owns a text generated by AI that was based on a writer’s work, or who owns the computer-generated likeness of an actor, are sparking heated debates.

More examples from various sectors of work are emerging every week. This raises the question: “Who should own what AI produces?”. Given the country’s strong cultural influence on this issue, the United States plays a major role.

How does AI work, how should it work, and how can humans deal with its consequences?

Artificial intelligence is a broad term describing various technologies, such as machine learning and natural language processing. What they all have in common is that they require data to produce an outcome. This data may come from the developers or AI companies themselves, or it may be obtained from the internet or other sources. This is where the basis for existing ownership issues lies. From an ethical perspective, everyone involved in the creation of a product should be properly credited. Furthermore, there is a widespread belief that a human is required to create art. US copyright law itself does not explicitly state who can be granted copyright, and thus ownership. It is therefore unclear whether AI could also be defined as an author and be protected by copyright laws, for example, when it creates content without human assistance. However, going beyond this initial definition, the US Copyright Office has gone on to state that authorship will only be granted if a human created the content.
Karl Marx has already addressed the question of ownership of the means of production. According to his theory, technologies such as AI could improve living conditions for workers and people in general, but they are often used to increase the value for the capitalist who owns them. Consequently, in a capitalist system, ownership lies with the capitalist. However, as an alternative, claims could be made by the workers training the AI (either directly or indirectly).

Expert Opinions

Experts on the subject – such as those working in the AI sector, academics, creators and policymakers – have expressed their views in numerous newspaper articles. A content analysis of these expert opinions suggests that AI can be seen as a new technology that creates opportunities for artists, but is limited by the way it operates. It relies on data created by humans to predict outputs, which can be problematic as this is frequently taken from the internet without regard for the ownership rights of the original creators.
Artists often do not give their consent for AI to be trained on their creations, and AI companies even breach copyright laws; this is why some experts argue that ownership should not lie with AI companies. Furthermore, the argument has been
put forward that the value of art and creative work is tied to humans creating intellectual property. When AI creates something, the individual using the AI still bears responsibility. However, remixing ideas has always been part of human history, and this concept could therefore also apply to AI practices. This is particularly important in a commercial context.
AI should enhance human creativity, yet it is currently being used to replace artists by taking over some of their work. In particular, legal experts regard transparency as essential for finding solutions. Regulators need to understand how AI works, and companies should make this more transparent. Only then will industry standards, guidelines or even government intervention be possible.
Thus, whilst few clear legal positions were expressed, one prominent aspect was that, as things stand, protected intellectual property can only be created by humans.

How those who create music, visual art or written content are facing these issues

Relevant cases that could be identified include AI creating visual art forms, AI composing music, AI generating text content, and AI regulation. Here, many similarities can be found, but also some differences. Whilst fears that AI will replace humans are expressed in all cases, the positive assessment varies. The emergence of AI is more frequently compared to past technological developments within the field of music than in other fields. Beyond these differences, the discussion surrounding the topic of AI and ownership focuses on similar aspects across all cases. Ethics took centre stage in this discussion, and the fear of humans being replaced by AI was raised several times, particularly in cases concerning the creation of art. Experts appear to be unanimous on how AI is changing the workplace. Linked to this, concerns about the origin of the data used to train these AI systems are also part of the discussion. A common view is that the output generated by AI based on this data still belongs to the original creators to some extent. Regarding the current regulatory landscape, experts across the various cases agreed that there is a lack of guidelines – a view shared even by the CEO of an AI company – whilst calling for regulations to be put in place. This was particularly evident in the case concerning regulation.

So what now? What should be done?

Currently, the problem lies in the fact that people creating content used by AI are not being properly compensated. This is where change needs to take place in the field of AI in order to
establish an ethical approach to working with this technology and prevent monopoly ownership by a single party. That is why, based on the analysis of expert opinions, the following next steps can be proposed:

• Voluntary submissions of data for AI training, rather than stealing art and infringing copyright
• Compensation and partial ownership for artists who create content used to train AI, to be provided by AI companies

What does the future hold?

The development and implementation of regulations governing AI ownership will be accelerated as the debate no longer focuses solely on ethical concerns, but will have direct economic consequences (as seen in the writers’ and actors’ strike). When regulations are inevitably put in place, AI systems will still wield considerable power, but may be viewed more positively by creators as a tool to expand their work. It remains to be seen who will ultimately be credited for this work under new regulations and standards. With regard to transparency, AI and its effects could be studied further by independent researchers to make its workings clearer to policymakers. Furthermore, a fair model of compensation for the various stakeholders could be established.

References

Acemoglu, D., & Johnson, S. (9 June 2023). Big Tech Is Bad. Big A.I. Will Be Worse. The New York Times. www.nytimes.com/2023/06/09/opinion/ai-big-tech-microsoft-google-duopoly.html

Barnes, B., Koblin, J. & Sperling, N. (13 July 2023). Actors Join Writers on
Strike, Bringing Hollywood to a Standstill. The New York Times. www.nytimes.com/2023/07/13/business/media/sag-aftra-writers-strike.h tml (last accessed on 2 August 2023)

Chan-Olmsted, S. M. (2019). A Review of Artificial Intelligence Adoptions in the Media Industry. International Journal on Media Management, 21(3–4), 193–215. doi.org/10.1080/14241277.2019.1695619

Coscarelli, J. (19 April 2023). An AI-generated hit featuring fake ‘Drake’ and ‘The Weeknd’
Rattles the Music World. The New York Times. www.nytimes.com/2023/04/19/arts/music/ai-drake-the-weeknd-fake.html

Draxler, F., Werner, A., Lehmann, F., Hoppe, M., Schmidt, A., Buschek, D. & Welsch, R. (2023). The AI Ghostwriter Effect: Users Do Not Perceive Ownership of AI-Generated Text But Self-Declare as Authors. In ACM, New York, USA. 1–34.

Elteren, M. V. (2003). U.S. Cultural Imperialism: Today Only a Chimera. SAIS Review, 23(2), 169–188. doi.org/10.1353/sais.2003.0038

Federal Trade Commission. (2023). Federal Trade Commission Act. FTC.
www.ftc.gov/legal-library/browse/statutes/federal-trade-commission-act
(last accessed on 27 July 2023)

Floridi, L. (2020). The Fight for Digital Sovereignty: What It Is, and Why It Matters, Especially for the EU. Philosophy & Technology, 33(3), 369–378. doi.org/10.1007/s13347-020-00423-6
Fouseki, K. & Vacharopoulou, K. (2013). Digital museum collections and social media: ethical considerations of ownership and use. Journal of Conservation and Museum Studies, 11(1), Article 5. 10.5334/jcms.1021209. doi.org/10.3390/math10152552

Fuchs, C. (2002). Social self-organisation in information society capitalism / 2 Crisis and critique in the information society: Essays on Herbert Marcuse, capitalist development and self-organisation. Books on Demand.

Next Move Strategy Consulting. (January 2023). Artificial intelligence (AI)
market size worldwide in 2021 with a forecast until 2030. NextMSC. www.nextmsc.com/report/artificial-intelligence-market (last accessed on 27 July 2023)

Rodgers, D. (14 July 2023). Shein’s secretive AI algorithm will steal your artwork. Dazed.
www.dazeddigital.com/fashion/article/60355/1/shein-mafia-barbie-ai-algorithm-will-steal-your-artwork-lawsuit-fashion-news (last accessed on 2 August 2023)

Mukhamediev, R. I., Popova, Y., Kuchin, Y., Zaitseva, E., Kalimoldayev, A., Symagulov, A., Levashenko, V., Abdoldina, F., Gopejenko, V., Yakunin, K., Muhamedijeva, E., & Yelis, M. (2022). Review of Artificial Intelligence and Machine Learning Technologies: Classification, Restrictions, Opportunities and Challenges. Mathematics, 10(15), 2552. doi.org/10.3390/math10152552

Nachtwey, O. & Staab, P. (December 2015). The vanguard of digital capitalism. eurozine, 21. www.researchgate.net/publication/292994969

OpenAI. (14 November 2022a). Sharing & publication policy
openai.com/policies/sharing-publication-policy h-the-openai-api-policy (last accessed on 27 July 2023) openai.com/blog/chatgpt

OpenAI. (30 November 2022b). Introducing ChatGPT.
openai.com/blog/chatgpt

Schlag, G. (2023). The European Union’s Regulation of Social Media: A Discourse Analysis of the Digital Services Act. Politics and Governance, 11(3). doi.org/10.17645/pag.v11i3.6735

Taulli, T. (2019). Artificial Intelligence Basics: A Non-Technical Introduction.
Apress. doi.org/10.1007/978-1-4842-5028-0

Turvill, W. (24 June 2022). Top 25 US newspaper circulations: Print sales fall another 12% in 2022. Press Gazette. pressgazette.co.uk/news/us-newspaper-circulations-2022/ (last accessed on 7 July 2023)

United States Copyright Office. (2021). Compendium of U.S. Copyright Office practices § 101. 3rd ed. 2021. Copyright. copyright.gov/comp3/chap300/ch300-copyrightable-authorship.pdf (last accessed on 26 July 2023).

United States Department of State. (2023). Intellectual Property Enforcement. State. www.state.gov/intellectual-property-enforcement/ (last accessed on 27 July 2023)

United States Patent and Trademark Office. (2023). Trademark, patent, or copyright. USPTO.
www.uspto.gov/trademarks/basics/trademark-patent-copyright (last accessed on 27 July 2023)
Fuchs, C. (2020). Communication and Capitalism: A Critical Theory. University of Westminster Press. doi.org/10.16997/book45

Veblen, T. (1898). The beginnings of ownership. American Journal of Sociology, 4(3), 352–356. doi.org/10.1086/210805